Executive summary of the meeting of the 2023 ASHP Commission on Goals: Optimizing Medication Therapy Through Advanced Analytics and Data-Driven Healthcare

artificial intelligence, clinical decision support, Commission on Goals, comprehensive medication management, data governance, medication-use process

Hybrid Live/Virtual Meeting

March 7, 2023

Address correspondence to Bruce Hawkins ([email protected])

Overview of the Commission

Each year ASHP convenes a Commission on Goals, composed of thought leaders in healthcare and related fields, to review societal and healthcare trends and developments that may affect ASHP members and the patients they serve and to provide guidance to the ASHP Board of Directors about potential strategic areas of focus for ASHP.^1-5 The March 7, 2023, Commission meeting focused on optimizing medication therapy through advanced analytics and data-driven healthcare. Members of the Commission ( appendix) were selected from key leaders in pharmacy, medicine, nursing, information science, government, other healthcare associations, and academia for their unique ability to discuss the use of advanced data analytics in healthcare. This publication is intended as a general overview of the discussion among participants and does not represent the official position of any of the individuals or organizations involved.

Controlling the weather

One Commission member likened managing data in healthcare to controlling the weather. As in other areas of today’s society, healthcare providers are encountering a deluge of data. The burden an increasing volume of data has on healthcare providers has been cited as a cause of provider burnout and negative patient outcomes.⁶ Commission members recognized that at its core, patient care is translating data into knowledge and then into action to care for patients. In the past, one of the primary challenges of patient care has been obtaining the data needed to provide the best care. One of the defining challenges of our time is divining from an overwhelming flow of information those data that are most important and using them to take appropriate action. Commission members were optimistic about the possibilities of harnessing the power of data to improve medication use but recognized the challenges and dangers healthcare leaders and providers face in accomplishing that task.

Data stewardship and governance

To successfully apply data to patient care, healthcare organizations need to put in place a structure to govern data use, a process some Commission members termed “data stewardship” to reflect the covenantal relationship created when organizations are entrusted with such sensitive data. Some Commission members suggested such a governance structure could find a parallel in pharmacy and therapeutics (P&T) committees, in which an organization identifies stakeholders and empowers them to institute policies that further the organization’s missions (eg, patient care, research, community service). Some Commission members suggested that, like P&T committees, data governance bodies should report to the organization’s chief medical officer. Maintaining a distinction from the organization’s information technology department and avoiding reporting to a chief operating or executive officer would enhance the data governance organization’s focus on clinical rather than operational aspects of its stewardship mission.

Commission members also suggested that successful data stewardship requires organizations to develop a data strategy. Although the strategies of organizations will differ based on their unique characteristics, Commission members did outline some common strategic priorities: enhancing the quality of patient care, including medication safety; translating population health data into patient care; ameliorating health inequities; preventing diversion of controlled substances; and ensuring data security and privacy.

Enhancing the quality of patient care

Commission members described the many ways data could be used to enhance patient care and outlined some of the challenges in accessing and translating that data into knowledge and action. Healthcare organizations will be challenged to integrate a variety of internal and external data sources (eg, payer and claims databases; patient information from remote monitoring systems and collected in patients’ own devices; and internally generated information from technology such as the electronic health record [EHR], smart pumps, and automated dispensing cabinets) to optimize patient care. Healthcare providers will be challenged to incorporate that data into action unless it is easily available. The dangers created when providers experience alert fatigue have long been known,⁷ so all this data needs to be translated into easily useable tools tailored to specific clinical interactions.

One of the prominent challenges in translating data into care is the lack of information system interoperability. The healthcare community’s experience with the development of the EHR was mentioned as an example of what can occur when interoperability is not incorporated as a core principle.³ One Commission member noted the evolution of the Office of the National Coordinator for Health Information Technology (ONC) Trusted Exchange Framework and Common Agreement (TEFCA) for health information networks (HINs).⁸ ONC’s goal in promulgating the TEFCA is to establish an “infrastructure model and governing approach for users in different networks to securely share basic clinical information” by setting out “a common set of non-binding, foundational principles for trust policies and practices that can help facilitate exchange among HINs.”⁸ The Commission member suggested that all healthcare stakeholders, including pharmacists, need to be involved in developing those standards and in similar future efforts, to avoid a repeat of the EHR experience and to make the standards most useful.

Supporting comprehensive medication management (CMM).

Commission members discussed ways in which data can be used to support CMM, which has been defined as “the standard of care that ensures each patient’s medications (ie, prescription, nonprescription, alternative, and traditional medications and vitamins or nutritional supplements) are individually assessed to determine that each medication is appropriate for the patient, effective for the medical condition, safe given the comorbidities and other medications being taken, and able to be taken by the patient as intended.”^9-11 CMM is an interprofessional team-based approach to optimizing medication therapy, and a central feature of CMM is having a pharmacist develop a plan with the patient or caregiver to optimize medication therapy, working in collaboration with the physician and other team members. Four pillars of CMM have been described, 2 of which rely on easy access to data:

1. Practice and care delivery transformation.

2. Use of health information technology to support that activity (eg, making clinical information available at the point of care to determine whether individuals have achieved the clinical goals of therapy).

3. Use of advanced and complementary diagnostic tools (eg, integrating pharmacogenomic testing results to reduce cost and improve patient outcomes).

4. Advancing the necessary policy reforms to promote CMM and fostering pilot programs and demonstration projects for those reforms.¹¹

Commission members expressed concern about the lack or inconsistency of data to measure the success of CMM, particularly for patients outside of hospitals and ambulatory care settings. Commission members noted laudable advances in capturing patient data through remote monitoring of wearables such as continuous glucose monitors and expressed optimism that similar technologies may come online for other therapies. It was recognized, however, that for most patients the sole source of data regarding medication adherence is claims data, which demonstrate only whether a prescription is filled and is not a measure of therapeutic benefit.

Developing, refining, and implementing quality measures.

Although Commission members noted the value of patient care quality measures, they shared concerns about meeting a seemingly ever-increasing number of measures, uncertainty about the relative importance of various measures, and the utility of some measures across varied healthcare settings. The Commission contemplated the possibility that artificial intelligence (AI)–based tools could improve quality measures by facilitating data capture and helping clinicians identify, evaluate, and refine quality measures. One Commission member provided an example of crowdsourcing weight-based dosing ranges in clinical decision support tools based on data captured from the EHR, particularly when prescribers routinely bypass alerts. While such a project would require rigorous clinical and legal risk review, it could relieve clinicians’ alert fatigue and result in improved patient care.

AI-based prediction models.

Commission members were provided an overview of several AI-based disease outcome prediction models that highlighted their strengths and limitations. Commission members noted that all the data sets used to develop AI models contain flaws that can create misleading results, from treatment effects to biases, including selection bias. Given these limitations, the process for deploying such a model into clinical use requires that clinicians and information specialists evaluate the model’s inputs and outputs not only during data preparation, model development, and model validation but also during its impact assessment and implementation into practice,¹² perhaps through a pragmatic clinical trial or similar process.¹³ Clinicians, informaticists, and software developers also need to keep in mind that their implementation data test set may have some of the same flaws as the training set used to construct the models, indicating a need for extended vigilance.

Commission members also noted the distinction between AI models based on tabular data (eg, clinical values such as patient outcomes or laboratory values) and deep learning models based on AI analysis of images or language. Over the past 5 years, models built on tabular data have become almost universally explainable through a variety of methods, suggesting that clinicians should require an explanation for a model before deployment, insisting on “glass box” rather than black box models.¹⁴ The Commission considered whether it would now be possible to mandate model explainability through law or regulation. It was noted that transparency was one of the 7 requirements for trustworthy AI developed by the European Commission High-Level Expert Group on Artificial Intelligence.¹⁵ In addition, Commission members noted that there are likely to be significant differences across the healthcare community in the level of trust regarding AI-generated models, data sets, recommendations, and even areas of EHR integration, including summarizing of notes. One Commission member suggested that the trust-building process for AI tools has a parallel in the process pharmacists have engaged in to demonstrate their value to fellow clinicians, and that pharmacists could make good champions for well-validated AI tools.

The impact of AI on healthcare and pharmacy practice was the subject of the 2019 meeting of the Commission on Goals; the report of that meeting provides more detailed information on the subject.²

Workforce preparation

Preparing the healthcare workforce for a digital future was the subject of the 2020 Commission on Goals meeting.³ As Commission members examined the efforts needed to prepare the healthcare workforce for a more data-driven practice, the discussion covered many of the same themes: the need to prepare learners to be able to adapt to practice changes over the course of their multi-decade careers, nurturing lifelong competencies while imparting specific practical skills; the challenges of reaching a diverse and ever-increasing number of audiences for learning; and how to harness the very technology changing healthcare to improve learner and practitioner instruction. One suggestion that garnered a great deal of discussion was the urgent need to collaboratively develop interprofessional competencies, given the increasingly team-based nature of healthcare. Another topic of discussion was the need to develop healthcare informatics specialists as well as informatics-trained clinicians to bridge the gaps between healthcare data, information technology, and patient care. The Healthcare Information and Management Systems Society Technology Informatics Guiding Education Reform (TIGER) Community initiative, a networking effort that began with nursing students but has since grown to include other disciplines, was cited as one model.¹⁶

Translating population health data into patient care

One of the promises and challenges of the rising use of data is translating population health data into patient care. One study ranked the relative contribution of determinants of health to overall health as follows: behavior, 40%; genetics, 30%; social factors 15%; healthcare, 10%; and environmental factors, 5%.¹⁷ As one author has pointed out, “[e]ven if we get the part that healthcare practitioners typically focus on completely right, that still leaves 90% of what determines health unaccounted for.”¹⁸ Commission members acknowledged the importance of incorporating social determinants of health, something perhaps as simple as a ZIP code, into patient care but recognized some of the challenges. Among those challenges is the veracity of the data used. One Commission member provided the example of a recent Office of the Inspector General study of Centers for Medicare and Medicaid Services data that found that inaccuracies in Medicare’s race and ethnicity enrollment data hinder our ability to assess health disparities.¹⁹

That discussion highlighted a growing data-use challenge: how to best utilize patient-generated data. Commission members noted that a distinction could be made between data generated by a Food and Drug Administration–approved medical device and less rigorously tested consumer products and suggested that there may be a role for regulators or healthcare associations in developing or vetting such consumer products to assist their member practitioners.

Addressing health inequities

Another promising aspect of the increasing use of data in healthcare is the capacity to alleviate or mitigate health inequities. Commission members discussed the importance of instilling the concept of “health equity by design” into data governance strategies.²⁰ Strategies will be needed to include populations that have previously been marginalized or whose historic lack of access to healthcare or technology may limit the data needed to provide optimal care. One example cited by a Commission member is the National Institutes of Health All of Us Research Program, which seeks to compile data from 1 million people of diverse backgrounds across the US to build a database to help determine how biology, lifestyle, and environment affect health and to discover ways to treat and prevent disease.²¹

Preventing diversion of controlled substances

Commission members discussed the utility of data analytics and AI in preventing diversion of controlled substances. The ongoing opioid epidemic and several high-profile prosecutions of manufacturers and prescribers have increased focus on diversion of controlled substances.²² Commission members acknowledged that funding for diversion prevention may sometimes lag other organizational priorities if it is narrowly viewed through a revenue lens, as it is revenue protective (eg, from institutional liability) rather than revenue generating. Given the liability and public health implications, health systems are beginning to embrace the use of data analytics and AI in preventing diversion of controlled substances.^23,24

Ensuring data security and privacy

Commission members discussed the challenges of keeping data secure and protecting the privacy of patient information. They acknowledged the competing demands of making data useful and accessible while protecting it from malicious use. Security breaches, including ransomware attacks, have become so common that one Commission member suggested it was a best practice to assume an intruder is already in your information system and to plan accordingly.

Commission members emphasized not only the legal obligation but also the moral imperative of maintaining the privacy of patient information. Commission members also discussed the increasing pressure from outside organizations (eg, specialty pharmacies) to share patient information as a condition of conducting business, which was described as essentially holding patient information hostage. Commission members noted that HIPAA is approaching 30 years of age and that the intervening changes in technology and culture require updated approaches. One Commission member expressed the fear that if the US does not address privacy issues, not just in healthcare but more broadly, we will find ourselves bound to standards developed in other parts of the world; the member cited the European Union’s General Data Protection Regulation, which is binding on any organization that targets or collects data related to people in the European Union.

Trends and opportunities identified by the Commission

Each Commission member was asked to identify at least one major trend or opportunity that, from their unique vantage point, they believe will have a significant impact on healthcare over the next 3 to 5 years, whether related to this year’s topic or not. One common theme was that the healthcare worker shortage will not be resolved quickly or easily and that healthcare will face difficult choices in the years to come, particularly in the face of continued financial constraints. The healthcare worker shortage was viewed as one of the factors driving an increasing level of interdependence of healthcare team members as well as a potential driver of more data-driven healthcare.

Some Commission members foresaw positive developments in population-based patient care, including a public health data modernization push that will help integrate public health data into clinical care, and increased population-level genetic health risk scoring and screening becoming more common. A caution was shared, however. In the context of whole-genome research and increasing use of genetic data in healthcare, genetic data privacy protection will become a significant challenge, and personalized medicine (eg, pharmacogenomics, digital health) may widen current health disparities unless the healthcare system adopts an equity-by-design culture. Some Commission members noted a need for national standards for data privacy and national licensing requirements for healthcare practitioners as they practice more across state lines.

One Commission member suggested that patients will continue to demand more telehealth options. Another Commission member pointed out that the pandemic had another, more dangerous outcome—disinformation (eg, vaccine conspiracy theories)—and suggested that the healthcare community will need to develop effective means of responding. It was noted that the pandemic resulted in increased data sharing across the healthcare continuum, and it was suggested that such sharing will only continue to grow. One Commission member predicted that ambient listening systems that capture and analyze provider-patient discussion to document patient encounters will come into wider use and reduce documentation burden, standardize documentation, and provide real-time clinical decision support.

Another Commission member suggested that the tension among the pharmaceutical industry, pharmacy benefit managers, and health systems regarding use of patient data may create the need for a neutral third-party arbiter regarding data use.

One Commission member noted the large number of healthcare quality measures and predicted that researchers and clinicians would focus more on streamlining them, aligning them across the continuum of care, and figuring out ways to evaluate their role in improving patient care. Another Commission member emphasized the need for a medication management data initiative to convince healthcare executives, the public, legislators, and regulators of the value of CMM.

Conclusion

The Commission on Goals: Optimizing Medication Therapy Through Advanced Analytics and Data-Driven Healthcare identified important trends created by the burgeoning use of data to deliver and measure healthcare. Commission members discussed the tremendous potential for data to improve patient care as well as some dangers and possible pitfalls. The Commission outlined 4 broad areas of opportunity for ASHP and other stakeholders in advancing the quality of data-driven healthcare care (resources, education, research, and advocacy) and briefly described other major trends or opportunities that may have a significant impact on healthcare over the next 3 to 5 years.

Acknowledgments

ASHP and the author gratefully acknowledge the following individuals for reviewing drafts of this document: David F. Chen, BSPharm, MBA; Sophia R. Chhay, PharmD; Daniel J. Cobaugh, PharmD, DABAT, FAACT; Joshua King, BA; Bayli Larson, PharmD, MS, BCPS; Sajel Lewis, PharmD, BCPS; Taylor Lindsay, PharmD, BCPS; Autumn Pinard, PharmD, MBA; Douglas J. Scheckelhoff, MS, FASHP; and Hannah K. Vanderpool, PharmD, MA.

Disclosures

The author has declared no potential conflicts of interest.

Appendix—Roster of 2023 ASHP Commission on Goals: Optimizing Medication Therapy Through Advanced Analytics and Data-Driven Healthcare

Thomas J. Johnson, PharmD, MBA, BCCCP, BCPS, FASHP, FCCM, FACHE

Commission Chair

Vice President of Hospital Pharmacy and Laboratory Services

Avera Health

Linda S. Tyler, PharmD, FASHP

Commission Vice Chair

Immediate Past President, American Society of Health-System Pharmacists

Professor (Clinical), Department of Pharmacotherapy, College of Pharmacy

University of Utah Health

Paul W. Abramowitz, PharmD, ScD (Hon), FASHP

Chief Executive Officer

American Society of Health-System Pharmacists

COMMISSION PARTICIPANTS

Amanda Brummel, PharmD, BCACP

Vice President Clinical Pharmacy Services

Fairview Health Services

Rich Caruana, PhD

Senior Researcher

Microsoft

Kathy Chappell, PhD, RN, FNAP, FAAN

Senior Vice President of Certification, Accreditation, Research, and Quality

American Nursing Credentialing Center

Micah Cost, PharmD, MS, CAE

Chief Executive Officer

Pharmacy Quality Alliance

Christopher A. Hatwig, RPh, MS, FASHP

President

Apexus LLC

Todd R. Henderson, PharmD

Lead Product Manager, Inpatient Pharmacy

Oracle Health & AI

John Komenda, PharmD

Software Developer

Epic

Tom Leary, MA, CAE, FHIMSS

Senior Vice President and Head of Government Relations

Healthcare Information and Management Systems Society (HIMSS)

Patsy McNeil, MD

System Chief Medical Officer

Adventist HealthCare

Stephen C. Mullenix, BSPharm, RPh

Senior Vice President, Public Policy & Industry Relations

National Council for Prescription Drug Programs (NCPDP)

Ken Perez, MBA

Vice President, Healthcare Policy and Government Affairs

Omnicell

Chris Rochon, PharmD

Senior Manager, Operations Compliance

Amazon PillPack

Scott M. Rochowiak, RPh

Manager, Product Performance

Surescripts

Shannon Sims, MD PhD

Senior Vice President, Data Operations

Vizient, Inc.

Lisa S. Stump, BSPharm, MS, RPh, FASHP

Senior Vice President, Chief Information and Digital Transformation Officer

Yale New Haven Health and Yale School of Medicine

Sara L. Van Driest, MD, PhD

Director of Pediatrics, All of Us Research Program

National Institutes of Health

Kasey K. Thompson, PharmD, MS, MBA

Commission Secretary

Chief Operating Officer & Senior Vice President

American Society of Health-System Pharmacists

COMMISSION ON GOALS OBSERVERS

Paul C. Walker, PharmD, FASHP

President, American Society of Health-System Pharmacists

Clinical Professor and Assistant Dean of Experiential Education and Community Engagement

University of Michigan College of Pharmacy

Manager

Michigan Medicine Department of Pharmacy

Nishaminy (Nish) Kasbekar, PharmD, BSPharm, FASHP

President-elect, American Society of Health-System Pharmacists

Chief Pharmacy Officer

Penn Presbyterian Medical Center

Christene M. Jolowsky, BSPharm, MS, FASHP, FMSHP

Treasurer, American Society of Health-System Pharmacists

System Pharmacy Director

Hennepin Healthcare